Image analysis – Applications
Reexamination Certificate
2000-09-07
2004-08-03
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
C348S463000
Reexamination Certificate
active
06771795
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to image processing, and, more particularly, to a channel for data transmission in, or watermarking of, video signals.
2. Description of the Related Art
Image information is commonly recorded or stored in either analog or digital format, but ultimately any information recorded in an analog format may be translated into a digital format. Once in digital format, such information is easily copied, modified, or distributed through electronic media such as the Internet. While protection by copyright is available to owners of the information once fixed in tangible media, theft through unauthorized copying, modification, or distribution of the information in digital format is easily accomplished and is difficult to detect from the information itself. For example, theft of compressed digital content (e.g., digital video disk (DVD), digital broadcast, or digital cinema) is a serious concern to owners of, and providers authorized to distribute, the compressed digital content. While the nature of the digital format allows for “perfect” copies of the compressed digital content, techniques exist to hide information within the compressed digital content that may be used to prove rightful ownership or use, such as in a court of law.
Collectively, these techniques are within the field of data verification, and for encoded and/or compressed digital image content, digital watermarking is commonly employed to insert hidden information into the content directly into the image video signal. Many watermarking techniques for digital images have been developed that allow content providers to prove the origin of their material and that their material was copied, modified, or distributed. In addition, published techniques exist for inserting invisible watermarks into video data, and the Copy Protection Technical Working Group (CPTWG) of the DVD forum, for instance, may select a standardized digital watermarking technique for DVD's.
For some systems, a separate data channel superimposed on the transmitted video signal may aid in data verification. The data channel may be added within the encoded packet stream itself or sent separately in secondary channels used for, for example, system, signaling, and synchronization information. In either case, the channel may be used to carry data corresponding to an encryption key, watermark pattern, or other form of watermark information.
In many of the techniques for encoded or compressed digital images, the watermark is “stamped” into the image by modifying or otherwise performing operations on either pixels or selected bit values of the pixels (termed the “pixel domain”). However, since these techniques work directly on the signal (e.g., compressed digital image) itself, they may either 1) distort the image, 2) be rendered useless by “perfect” copying, or 3) be lost by legitimate signal processing operations during transport of the image (e.g., compressed data transport or storage). Techniques for inserting watermarks in the pixel domain are generally not robust to common image signal processing methods, such as scaling, cropping, rotation, and compression, of the digital image information. Furthermore, techniques designed to provide higher bit rates for hidden data signals of the watermark tend to be more visible in the watermarked image and/or less robust to subsequent signal processing.
For example, some pixel-domain watermarking techniques exploit a human's insensitivity to high spatial and/or temporal frequencies, and insert low-level watermarks with these characteristics. Unfortunately, the watermarks of these techniques tend to be degraded or obliterated completely when the image is resized (specifically, reduced in size) or compressed by commonly used algorithms such as those specified in the MPEG-1 or MPEG-2 standards. Furthermore, the rapid variations associated with the watermark are often interpreted as noise by signal processing algorithms encoding or compressing the image data, which tends to lower the efficiency of compression algorithms. Other pixel-domain watermarking techniques specifically target edges of objects in the image as a means for inserting a watermark or hiding data. However, since data compression algorithms such as those employed in MPEG-2 systems introduce noise at object edges, a similar degradation or obliteration of the watermark may occur.
SUMMARY OF THE INVENTION
The present invention relates to a channel allowing for watermarking of an image or data transmission within an image that is robust to subsequent signal processing methods, such as scaling, cropping, rotation, and compression. The presence of the channel itself may be considered as the watermark, or the watermark may be represented by data carried in the channel. The channel is inserted into the image by varying display characteristics of the image across at least one or more regions of the display in accordance with a known function. Variation of a display characteristic for preferred embodiments may be a spatial variation that varies with time based on a known spatio-temporal function. Examples of such display characteristics are center-to-side 1) brightness falloff and 2) resolution reduction. For most display devices, although attempts may be made to reduce or eliminate display variations, human viewers may tolerate and accept these subtle variations in a displayed image. For exemplary embodiments in accordance with the present invention, slightly changing one or more display characteristics from center-to-side in accordance with a predefined function allows for the channel within the image since viewers may not be aware that the display characteristics are changing.
In accordance with one embodiment of the present invention, watermark information is encoded into a video signal. The watermark information is encoded by: (a) applying a spatial function to one or more components of the video signal to vary a display characteristic over at least one region of a frame of the video signal; (b) varying the spatial function based on the watermark information; and (c) repeating (a) and (b) for one or more additional frames of the video signal to encode the watermark information into the video signal.
In accordance with another embodiment of the present invention, watermark information in a video signal is detected. The watermark information is detected by: (a) generating a measure between frames of the video signal for one or more components in one or more regions; (b) determining whether one or more regions are still based on the generated measure; and (c) detecting a variation in a sub-set of the components for each region that is still, the variation representing the watermark information.
REFERENCES:
patent: 6094722 (2000-07-01), Astola et al.
patent: 6211919 (2001-04-01), Zink et al.
patent: 6381367 (2002-04-01), Ryan
patent: 6594373 (2003-07-01), Gustafson
patent: 6636638 (2003-10-01), Yamaguchi
patent: WO 96/06503 (1996-02-01), None
Szepanski, “Binärdatenübertragung über Videokanäle mit Datensignalen sehr geringer Amplitude,”FERNSEH- und KINO- TECHNIK,vol. 32, No. 7, 1978, pp 251-256 (with English translation).*
Szepanski, “Additive Binärdatenübertragung für Videosignale,”NTG-Fachberichte,vol. 74, 1980, pp. 343-351 (with English translation).*
Dittmann et al., “A new approach for transformation invariant image and video watermarking in the spatial domain: SSP—self spanning patterns,”Proc. SPIE vol. 3971: Security and Watermarking of Multimedia Contents II,Jan. 2000, pp. 176-185.
Burke, Esq. William J.
Johns Andrew W.
Sarnoff Corporation
LandOfFree
Spatio-temporal channel for image watermarks or data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Spatio-temporal channel for image watermarks or data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spatio-temporal channel for image watermarks or data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3294327